X. Frequency 101. While converters can be designed for any frequency, the use of high frequency, as 60 cycles, imposes more severe limita- tions on the design, especially that of the commutator, as to make the high-frequency converter inferior to the low-frequency or 25-cycle converter. The commutator surface moves the distance from brush to next brush, or the commutator pitch, during one-half cycle, that is, 3^50 second with a 25-cycle, J^20 second with a 60-cycle converter. The peripheral speed of the commutator, however, is limited by mechanical, electrical, and thermal considera- tions— centrifugal forces, loss of power by brush friction, and heating caused thereby. The limitation of peripheral speed limits the commutator pitch. Within this pitch must be in- cluded as many commutator segments as necessary to take care of the voltage from brush to brush, and these segments must have a width sufficient for mechanical strength. With the smaller pitch required for high frequency, this may become impossible, and the limits of conservative design thus may have to be exceeded. In a converter, due to the absence of armature reaction and field distortion, a higher voltage per commutator segment can be 258 ELEMENTS OF ELECTRICAL ENGINEERING . allowed than in a direct-current generator. Assuming 17 volts as limit of conservative design would give for a 600-volt con- verter 36 segments from brush to brush. Allowing 0.2 inch for segment and insulation, as minimum conservative value, 37 segments give a pitch of 7.4 inches. Estimating 5000 feet per minute as conservative limit of commutator speed gives 83.3 feet or 1000 inches peripheral speed per second, and with 7.4 inches pitch this gives 136 half cycles, or 68 cycles, as limit of the frequency, permitting conservative commutator design. At 60 cycles higher voltage per segment, narrower segments and higher commutator speeds thus are necessary than at 25 cycles, and the 60-cycle converter, though still within conserva- tive limits, does not permit as conservative commutator design, especially at higher voltage, as a low-frequency converter, and a lower self-inductance of commutation thus must be aimed at than permissible in a 25-cycle converter, the more so as the fre- quency of commutation (half the number of commutator seg- ments per pole times frequency of rotation) necessarily is higher in the 60-cycle converter. . . Somewhat similar considerations also apply to the armature construction : the peripheral speed of the armature, even if chosen higher for the 60-cycle converter, limits the pitch per pole at the armature circumference, and thereby the ampere conductors per pole and thus the armature reaction, the more so as shallower slots are necessary. The 60-cycle converter cannot be built with anything like the same armature reaction as is feasible at lower frequency. On the armature reaction, however, very largely depends the stability of a synchronous motor or converter, and machines of low armature reaction tend far more to surging and pulsation of current and voltage than machines of high armature reaction. The 60-cycle converter therefore cannot be made quite as stable and capable of taking care of violent fluctuations of load and of excessive overloads as 25-cycle converters can, and in this respect the lower-frequency machine is preferable, though under reasonably favorable conditions regarding variations of load, variations of supply voltage, and overload 60-cycle con- verters give excellent service. It is this inherent inferiority of the 60-cycle converter which has largely been instrumental in introducing 25 cycles as the frequency of electric power generation and distribution. SYNCHRONOUS CONVERTERS 259 At 25 cycles, converters are used on railway load — the most fluctuating and therefore most severe service — built for 1200 volts, and even still much higher voltages are available.